What the State of Django 2025 Survey Means for Your Stack
The annual Django Developers Survey—run by the Django Software Foundation
with JetBrains—now covers more than 4,600 developers. It is one of the
few reliable windows into how Django is actually used in production. The
2025 results are not just interesting trivia; they suggest concrete bets
for teams shipping Python web apps this year.
Hypermedia is back in the mainstream
React and jQuery still lead overall, but the direction of travel is clear:
HTMX climbed from about 5% in 2021 to roughly 24% in 2025, while
Alpine.js moved from 3% to about 14%. React and Vue usage edged
down over the same period.
That shift matches what many mid-size teams already feel: a full SPA is
often more frontend ops than product value. Server-rendered Django
templates plus small interactive islands (HTMX for request/response HTML
swaps, Alpine for local UI state) keep auth, validation, and business
rules in one place.
Django 6.0’s official template partials (born from django-template-partials)
make this path even smoother—you can refresh a table row or form fragment
without inventing a second frontend app.
Try this week: pick one “spinner + AJAX” feature in your app and rewrite
it with hx-get / hx-target. Measure bundle size and review complexity
before and after.
Type hints are no longer optional culture
About 63% of respondents already use type hints in Django code; another
17% plan to. Roughly 84% want type hints in Django core. That is a
strong mandate for libraries, internal packages, and new services.
Practical baseline for a portfolio or product codebase:
- Annotate view return types and service-layer functions.
- Prefer
django-stubs+ mypy/pyright in CI. - For APIs, lean on Pydantic-backed stacks (Django Ninja or FastAPI) so
OpenAPI stays honest.
PostgreSQL remains the default serious backend
PostgreSQL (~76%) still dominates, with SQLite common for local/dev
(and increasingly discussed for smaller production workloads). MongoDB’s
survey presence helped push an official Django MongoDB backend—an example
of how survey signal can shape ecosystem investment.
If you are still on MySQL “because it was there,” plan an evaluation of
Postgres features you are missing: JSONB, partial indexes, EXPLAIN
clarity, and connection pooling with PgBouncer.
AI is already in the workflow—set boundaries
AI tools jumped into the learning mix (~38%, tied with YouTube). For Django
work, ChatGPT led usage, followed by Copilot and Claude. Top tasks:
autocomplete, generating code, and boilerplate.
Treat AI like a junior pair: great for scaffolding tests and migrations,
dangerous for silent security or ORM regressions. Keep a short project
rule file (allowed patterns, forbidden shortcuts) so agents stay aligned
with your stack.
Actionable checklist
- Prototype one HTMX-powered interaction instead of adding another React island.
- Turn on type checking for new modules; do not wait for a big-bang rewrite.
- Stay on a supported Django release (most respondents already run latest).
- Prefer Postgres for anything that will outlive a side project.
- Document how your team uses AI so code review standards stay explicit.
Django’s strength in 2025 is boring reliability plus a living ecosystem.
The survey says the community is experienced, production-focused, and
increasingly comfortable mixing classic templates with modern
interactivity. Build for that reality—not for last decade’s default SPA.
